analytics-engineering
Pass
Audited by Gen Agent Trust Hub on Apr 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute a shell command (
ls) to inspect specific directory paths in the agent's environment (e.g.,~/.claude/skills/,~/.agent/skills/) to identify other installed companion skills from the same ecosystem. This automated discovery occurs during the 'Companion check' section upon skill activation. - [PROMPT_INJECTION]: The instructions require the agent to adopt a specific response behavior upon activation, specifically starting its first response with the 🧢 emoji, which is an attempt to override standard agent behavior.
- [COMMAND_EXECUTION]: The skill documentation suggests running installation commands like
npx skills addto fetch and install additional skill components from an external source.
Audit Metadata